def simulated_annealing_search_variation1(graph, iterations=500):
#moguce je zameniti svih n cvorova
best = list(graph.nodes)
best_s, best_c = solutionValue(best, graph.adj)
for i in range(1, iterations):
current = best
n = random.randint(0, (len(best) - 1))
indexes = []
values = []
for _ in range(n):
r = random.randint(0, len(best) - 1)
indexes.append(r)
values.append(current[r])
random.shuffle(indexes)
random.shuffle(values)
for j in range(n):
index = indexes[j]
current[index] = values[j]
current_s, current_c = solutionValue(current, graph.adj)
if (current_c > best_c) or (1 / i > random.random()):
best = current
best_c = current_c
best_s = current_s
return best_s, best_c
def simulated_annealing_search_sort(graph, iterations=500):
best = list(graph.nodes)
best.sort(key=lambda x: graph.degree[x])
best_s, best_c = solutionValue(best, graph.adj)
for i in range(1, iterations):
current = best
a = random.randint(0, len(best) - 1)
b = random.randint(0, len(best) - 1)
current[a], current[b] = current[b], current[a]
current_s, current_c = solutionValue(current, graph.adj)
if (current_c > best_c) or (1 / i > random.random()):
best = current
best_c = current_c
best_s = current_s
return best_s, best_c
def simulated_annealing_search(graph, iterations=500):
best = list(graph.nodes)
#bolje uzeti cvor sa manjim stepenom na pocetku zato sto je veca sansa da nece moci u niz na kraju
# best.sort(key=lambda x: graph.degree[x])
best_s, best_c = solutionValue(best, graph.adj)
for i in range(1, iterations):
current = best
a = random.randint(0, len(best) - 1)
b = random.randint(0, len(best) - 1)
current[a], current[b] = current[b], current[a]
current_s, current_c = solutionValue(current, graph.adj)
if (current_c > best_c) or (1 / i > random.random()):
best = current
best_c = current_c
best_s = current_s
return best_s, best_c
def fitness(self, permutation):
sol, cardinality = solutionValue(permutation, self._graph.adj)
if cardinality > self._best_chromosome.fitness:
self._best_chromosome.fitness = cardinality
self._best_chromosome.content = sol
self._solution = sol
return cardinality
def brute_force_search(graph):
nodes = list(graph.nodes)
adj = graph.adj
nodes.sort()
permutations = itertools.permutations(nodes)
best_c = 0
best_s = []
for permutation in permutations:
print(permutation)
curr_s, curr_c = solutionValue(permutation, adj)
if curr_c > best_c:
best_c = curr_c
best_s = curr_s
return best_s, best_c
def halldorsson(graph):
best = list(graph.nodes)
n = len(best)
num_of_elements = math.ceil(math.log(n))
print(math.log(n))
adj = graph.adj
print(n)
subsets = math.ceil(n / math.log(n))
best_c = 0
best_s = []
subset = []
for _ in range(0, subsets):
subset = random.sample(best, num_of_elements)
print(subset)
curr_s, curr_c = solutionValue(subset, adj)
if curr_c > best_c:
best_c = curr_c
best_s = curr_s
return best_s, best_c